Jakub Macina
Jakub Macina has 6 papers in our library, in 2022–2025, most of them on Reasoning, Language models & pretraining, Dialogue systems; often with Mrinmaya Sachan, Iryna Gurevych, Manu Kapur.
OpenAlex · ORCID · h-index 8 · 35 works · 196 citations OpenAlex, 2026-10-14
📄 6 works in our library · 🔗 Cited 216 times in our library · ⭐ Paper Stars ranking: #5399
Affiliations
Fields, topics and concepts
📚 Shelves 10The field of science this work belongs to (OpenAlex fields).Computer Science · 6 worksArts and Humanities · 4 worksArtificial Intelligence · 6 worksLanguage and Linguistics · 4 worksInformation Systems · 5 worksSocial Sciences · 1 workEducation · 1 workMedicine · 1 workDemography · 1 workHealth Informatics · 1 work
🧭 Topics 4Research subjects and languages the work is about.Reasoning · 4 worksDialogue systems · 2 worksLanguage models & pretraining · 3 worksEvaluation & benchmarks · 1 work
Works by year
2020s
Papers
- MathDial: A Dialogue Tutoring Dataset with Rich Pedagogical Properties Grounded in Math Reasoning Problems (2023 · 131 citations)
Research on automatic dialogue tutors has been limited by a shortage of large, pedagogically sound datasets. - Automatic Generation of Socratic Subquestions for Teaching Math Word Problems (2022 · 70 citations)
The paper investigates whether large language models can automatically produce Socratic-style guidance for math word problems (MWPs). - Opportunities and Challenges in Neural Dialog Tutoring (2023 · 51 citations)
Neural conversational systems built on large language models have advanced quickly, but dialog tutoring has seen little benefit from this progress. - Stepwise Verification and Remediation of Student Reasoning Errors with Large Language Model Tutors (2024 · 35 citations)
LLM-based math tutors can solve word problems well, but they often fail to notice where a student went wrong. - MathTutorBench: A Benchmark for Measuring Open-ended Pedagogical Capabilities of LLM Tutors (2025 · 33 citations)
Judging how well an LLM teaches , rather than how well it solves problems, has relied on two weak options. - From Problem-Solving to Teaching Problem-Solving: Aligning LLMs with Pedagogy using Reinforcement Learning (2025 · 19 citations)
LLMs tuned for answering questions make weak tutors because they tend to reveal solutions instead of guiding learners.
Co-authors here
- Mrinmaya Sachan ×6
- Iryna Gurevych ×5
- Manu Kapur ×5
- Nico Daheim ×5
- Tanmay Sinha ×3
- Ido Hakimi ×2
- Kumar Shridhar
- Lingzhi Wang
- Mennatallah El-Assady
- Sankalan Pal Chowdhury
Venues
- Conference on Empirical Methods in Natural Language Processing (EMNLP) ×4: EMNLP 2025 ×2, EMNLP 2024, EMNLP 2022
- Conference of the European Chapter of the Association for Computational Linguistics (EACL): EACL 2023
- Findings of the Association for Computational Linguistics: Findings of the Association for Computational Linguistics 2023
Cited by
Updated
This page in Markdown → https://md.research.elamur.ai/free/en/authors/jakub-macina/